日本語版

Research note 021 — 2026-09-01

A World Map of Debittering

Mediterranean olives, Japanese mountain greens (sansai), Andean potatoes and lupin beans, West African leaf vegetables, Southeast Asian bitter gourd. These plants grow in different places, are eaten within different food cultures, and are talked about in different languages, yet they share one thing in common: as they are, they are too bitter to eat, or carry a bitterness that is too strong. Rearranging the processing records held at NIGAMI LAB, several types come into view in the「抜き方」 (in translation: "way of drawing the bitterness out") used on it. Some rely on water, some freeze and press, some leave it to microbes, some rub and soak it — this note reopens, along the axis of geography, the "removal" quadrant set up in theprevious note, "Can Cooking Create Bitterness?".

What this page establishes

• Debittering processes recorded around the world fall into several types — soaking in water, freezing and pressing, leaving it to microbes, rubbing and soaking — each recorded in separate sources.
• For lupin beans (tarwi), one study reports that a process combining water treatment and boiling lowers seed alkaloids by 89–97%.
• No study has yet directly compared which type reduces bitterness the most.
• One field-demonstration study observes that bitterness does not always reach zero even after a labor-intensive process.

Relying on water — parboiling and soaking

The most widely recorded type puts water itself to work. In Japan, a cooking practice is recorded as a local dish of Shimane: fukinoto (butterbur sprouts) are boiled in water with salt and baking soda added, then soaked in water to remove egumi before being mixed into miso.

In the Andean highlands, lupin beans (tarwi), which contain bitter quinolizidine alkaloids, are traditionally soaked in salted water and washed. A process combining water treatment and boiling is reported to remove 89–97% of the alkaloids in the seed.

Southeast Asia also has a process recorded with the same underlying skeleton. Thai local-government materials record, for mara khinok — a small, strongly bitter variety of bitter gourd — a cooking practice of "boiling it in boiling water with salt added, and once cooked through, discarding the water to reduce bitterness." Read alongside Japan's parboil-and-discard aku removal, the skeleton of the process looks similar.

In Thailand, the same kind of process is also recorded for leaf vegetables. A Mahidol University recipe records that the bitter leaves and flowers of a Senna species are "boiled, the water discarded, boiled again until soft, and the liquid squeezed out," with the number of repetitions and the endpoint being specific to this recipe.

Freezing and pressing — physical dehydration in highlands and rainforest

There are also more labor-intensive types that don't stop at simply soaking in water. On the Peruvian altiplano (above 3800m elevation), potatoes — including native bitter varieties — are frozen multiple times, taking advantage of below-freezing night-time winter temperatures.

The frozen potatoes are then submerged for 3 weeks or more in flowing river water, inside an enclosure made of netting and stakes.

For the final peeling, specialized tools are recorded as being used, such as "taquiña", a netted cloth used with treading pressure. Freezing and river immersion change the potato's structure, and physical pressure is then applied to remove the skin and moisture — tunta is less a chemical debittering than a dehydration technique that combines climate and physics.

The same physical process of pressing also appears in Amazonian cassava processing. In indigenous traditional technique, the grated root pulp is passed through a woven, basket-like press called "tipiti", and pressure is applied from above to squeeze it. One point needs care, though: the material processed this way is called "bitter cassava (mandioca-brava)", but according to the records, this name derives from the high level of cyanogenic compounds in the root, and does not necessarily refer to bitterness as a taste.

For that reason, this process cannot simply be called bitterness removal itself. Even so, it is worth recording that the same physical process of pressing has been documented, in separate sources, from two contrasting environments — the sub-freezing Andes and the Amazon rainforest.

Leaving it to microbes — making time itself the process

Mediterranean table olives are a type that entrusts time to microbes and enzymes. Early in storage, enzymatic hydrolysis of oleuropein is recorded as proceeding, followed by a slower chemical hydrolysis.

This whole sequence of processing is recorded as the main family of methods that reduce the bitterness of table olives. However, how fast it comes out and how much bitterness ultimately remains vary with the method and the final product style.

Rubbing and soaking — the hand decides the endpoint

West African bitter leaf (Vernonia amygdalina) has yet another type recorded. Boiling, soaking with several changes of water, and rubbing it clean in cold or warm water — here it is the hand's own sense, not a tool or the climate, that decides the endpoint.

A survey covering the Igbo cultural area of southeastern Nigeria likewise records bitter leaf as "bitter, usually washed before use." This record, however, does not give an endpoint for where the washing should stop.

There's also a type that hides it, rather than removing it

Most of the types covered so far are processes oriented toward "removal" — trying to reduce the cause of bitterness itself (though, as noted above for freezing and pressing, that cannot simply be called removal itself). Separately from this, NIGAMI LAB also has records of a masking type: a way of changing only how bitterness is felt, without reducing its cause. In one Indian study, for example, adding β-cyclodextrin to bitter-gourd juice is reported to have raised acceptability by enveloping the bitter compounds.

This reported value, however, is a hedonic 「受容性」 (in translation: "acceptability") score from a sensory evaluation using juice; it does not measure the intensity of bitterness itself. Separately, an evaluation by a trained panel on an extract has also reported a reduction in bitterness intensity itself. Removal techniques and masking techniques share the same goal but differ in mechanism — this boundary continues directly from the「抑制」 (in translation: "suppression") quadrant covered inthe previous note.

It doesn't always disappear completely

And what emerges from gathering these "removal techniques" together is that removing it completely is not always guaranteed. The flavor of good-quality tunta is described as「心地よく、わずかに淡白」 (in translation: "pleasant, slightly bland"), but in several processing evaluations including the improved variety silver, a「ごくわずかで取るに足らない程度の苦味」 (in translation: "slight, practically insignificant bitterness") has also been reported as detected.

This is a record from a single small-scale field-demonstration study, and cannot be generalized. Even so, the fact that bitterness does not always reach zero even after the labor-intensive process of freezing, river immersion, and peeling is something worth stating honestly as the close of this world map.

An open question

The "removal techniques" laid out in this map were each picked up individually from separate sources. Does research exist that compares them cross-sectionally along the axes of process, vocabulary, and social positioning? Tunta, bitter cassava, and tarwi, for instance, are each turned into a staple food through a chain of pre-processing steps — freezing, pressing, water-leaching, fermentation — but can the three of them be compared as the same type of 「克服の苦味」 (in translation: "bitterness overcome")? This question is not yet at a stage where it can be written as a claim, so it is kept, as an editorial structural hypothesis, in thehypothesis collection box. This map will be updated once primary sources directly comparing these processes are found.

Frequently asked

How is the bitterness of tarwi (lupin beans) removed?

Lupin beans (tarwi) are recorded as traditionally having been soaked in salted water and washed, and one study reports that a process combining water treatment and boiling lowered the seed's alkaloids by 89–97%. Multiple processing methods are known — biological, chemical, aqueous — and processes finishing in as little as 1 hour have also been reported. However, this source does not let us trace the species or conditions of the individual studies behind the 89–97% figure. There are also reports that combining fermentation lowers it further, but the results vary widely depending on the strain and process conditions — a variation the research itself states is 「説明できない」(in translation: "something it cannot explain").

Why are frozen potatoes (chuño/tunta) soaked in a river?

In the highlands of the Puno region, a process is recorded in which potatoes are frozen multiple times under conditions where night-time temperatures drop below 5℃, and are then soaked for about 3 weeks in flowing river water inside an enclosure built of netting and stakes. A field-demonstration study also reports that the average soaking period was about 25 days. Combining freezing with soaking in flowing water is explained as a process that changes the potato's structure, leading into the subsequent peeling and drying — but this source contains no chemical analysis that directly measured how much this soaking reduces the potato's bitter compounds themselves.

What is not known

• The individual primary studies (species, conditions) behind the 89–97% figure for the reduction in lupin-bean alkaloids have not been traced from this source.
• No source has been found that chemically measures how much the Andean freezing-and-river-immersion process reduces the amount of the potato's bitter compounds themselves.
• For the preparation of Thai bitter gourd and leaf vegetables, no quantitative measurement has yet been made linking boiling time or water volume to the amount of bitterness reduction.
• For debittering by rubbing and washing, no source gives a standard defining the endpoint of how much washing is enough.